Alternative Immunosuppression in Acquired Haemophilia A

To the Editor,

International recommendations [1] suggest prednisolone with rituximab or cyclophosphamide for patients with acquired haemophilia A (AHA) and a factor VIII activity of <1 IU/dL or an inhibitor titre of >20 Bethesda units (BU/mL). The guidelines cite an infection risk in frail patients [1]. Obaji et al. [2] report normalisation of FVIII activity and an undetectable inhibitor following treatment with MMF in 5/7 patients. Successful use of azathioprine has also been reported in AHA [3-8]. We prefer MMF over cyclophosphamide due to the risk of severe infections [2]. Herein, we present retrospective data regarding the use of MMF in AHA in a series of 19 consecutive patients (Table 1) from 2014 to 2023, including the use of azathioprine in three patients following relapse.

Prednisolone (1 mg/kg/day) was given for 2 to 3 weeks followed by a taper of 10 mg per week to 10 mg; 5 mg for 1 week and then 5 mg on alternate days for 1 week. The dose of MMF started at 500 mg BD and was escalated to 1 g BD as tolerated, typically after 1 to 2 weeks. MMF was stopped after two consecutive FVIII activity results >50 iU/dL with an undetectable inhibitor titre.

12/19 (63.1 %) were male. The mean age at presentation was 70.9 years. Haemostatic treatment was with FEIBA (13/19, 68.4 %) or Obizur (4/19, 21.0 %) or both (1/19, 5.2 %). One patient did not require haemostatic treatment. The minimum time between two consecutive FVIII activity >50 iU/ dL prior to stopping MMF was 14 days. 8/19 patients (42 %) achieved complete normalisation of their FVIII activity. 7/8 patients with complete normalisation of their FVIII activity had low titre inhibitors (<20 BU/mL). The mean time to normalisation of FVIII activity in responders was 4.3 weeks (range 0.6–10.1 weeks).

3/19 (15.8 %) (patients 1, 2, 3) had a complete normalisation of their FVIII activity with no relapse after stopping MMF. 5/19 (26.3 %) (patients 4, 5, 6, 7, 8) initially normalised their FVIII activity but subsequently relapsed. Two of these (patients 4 and 8) relapsed whilst on MMF, at 24 and 19 weeks, respectively, necessitating a change in treatment. One patient (patient 5) had a late relapse more than 6 months after stopping MMF. Two patients (patients 6 and 7) had an early relapse less than 6 months after stopping MMF. Patient 6 restarted and later successfully discontinued MMF. 3/19 (15.8 %) (patients 9, 12, 13) had an improved but not normalised FVIII activity following initiation of MMF. This corresponded to an absence of bleeds and an improved FVIII activity sustained in the ‘mild’ range for between 4 and 20 weeks. 8/19 (42.1 %) (patients 10, 11, 14–19) had no response to MMF. 6/8 (75 %) non-responders had high presenting inhibitor titres >20 BU/mL.

Bleeds in people who normalised their FVIII activity on MMF included delayed bleeding following angiography from a radial artery puncture site, with a FVIII activity of 13 iU/dL, and blood-stained rectal discharge, with a FVIII activity over 50 IU/dL. In both patients, the bleeding was successfully treated with tranexamic acid. One patient had a thigh bleed at the original site of bruising 13 days after admission, with a FVIII activity over 5 iU/dL. Patient 5 experienced rectal and cutaneous bleeding, and patient 6 produced bloodstained sputum after MMF had been withdrawn. These non-major bleeds did not require haemostatic treatment and responded to re-initiation of MMF.

7/8 patients (87.5%) who normalised their FVIII activity appeared to have idiopathic AHA. Patient 4 had active giant cell arteritis which was treated with methylprednisolone in the days prior to their AHA diagnosis, a likely confounding factor. The patient had a normalisation of FVIII activity at 5.4 weeks.

Fourteen infections were documented in 8 patients receiving MMF; 8/14 infections were managed in the community. Five patients required dose reductions or discontinuation of MMF due to gastrointestinal side effects.

Three patients with relapsed AHA were subsequently given azathioprine 50 mg BD, escalating to 100 mg BD as tolerated, after confirmation of normal TPMT activity and liver function tests. Two of these patients had not responded to second-line rituximab but had a prolonged response to azathioprine with a normalised FVIII activity and undetectable inhibitor titre at 42 and 78 weeks, respectively. Drug discontinuation was achieved at 72 weeks and 90 weeks, respectively. One non-major bleed, that did not require haemostatic treatment, and four infections occurred.

Overall, 7/11 patients (63.6%) with low titre inhibitors (<20 BU/mL) at presentation normalised their FVIII activity level with first-line steroids and MMF. 1/8 patients (12.5%) with high titre inhibitors normalised their FVIII activity level with first-line steroids and MMF. Therefore, the combination of prednisolone and MMF appeared particularly effective in patients with low-risk AHA with a presenting Bethesda titre <20 BU/mL.

Generalisability is limited as this is a small retrospective data set. Whilst ideally, the goal of treatment of AHA would be to achieve rapid inhibitor eradication with discontinuation of immune suppression, in more elderly patients with comorbidities, an alternative goal of treatment would be to avoid bleeds and life-threatening infections. We note recent advances regarding the off-label use of emicizumab, a bispecific monoclonal antibody that mimics FVIII, in AHA. A phase II study [10] supports the use of emicizumab as bridging haemostatic prophylaxis whilst awaiting the onset of immune suppression. The use of emicizumab may be a useful adjunctive strategy whilst the FVIII activity is significantly reduced, when immune suppressive agents may take many months to work. Furthermore, where patients have a short life expectancy, emicizumab alone may have a role in preventing bleeding in lieu of immune suppression.

In conclusion, we found the combination of steroids and MMF to be effective in low-titre but not high-titre AHA. Based on our findings, we suggest using steroids and MMF as first-line therapy instead of more potent immunotherapy for patients at risk of adverse effects, such as frail or elderly patients. Azathioprine may be useful in specific circumstances. Patients who are not candidates for immunosuppression and have no significant bleeding symptoms may benefit from an approach of active monitoring.

Charles Percy, Jayna Mistry, Will Lester and Gillian Lowe were involved in patient care. Charles Percy, Will Lester and Gillian Lowe revised the manuscript. Jayna Mistry wrote the manuscript and subsequently revised it.

The authors are compliant with the Best Practice Guidelines on Research Integrity and Publishing Ethics. Note consent has not been obtained from each case described in the table as the data was originally anonymised for age and gender.

The authors have nothing to report.

There are no conflicts of interest. J.M. has worked in an out of programme–research post funded by BioMarin as a Haemophilia Research Fellow. G.L. has received honoraria for participating in educational events from Novartis, Leo, Alexion, Sobi, Takeda, Novo Nordisk and Sanofi. She is the recipient of research funding (fellow post) from BioMarin. W.L. has received honoraria (including speaker fees, travel grant, advisory board) from Takeda, Octapharma and CSL Behring. C.P. has received honoraria from LFB, Takeda, Novo Nordisk, Sobi, Roche.